CN110686679B - High-orbit optical satellite offshore target interruption track correlation method - Google Patents

High-orbit optical satellite offshore target interruption track correlation method Download PDF

Info

Publication number
CN110686679B
CN110686679B CN201911036959.XA CN201911036959A CN110686679B CN 110686679 B CN110686679 B CN 110686679B CN 201911036959 A CN201911036959 A CN 201911036959A CN 110686679 B CN110686679 B CN 110686679B
Authority
CN
China
Prior art keywords
track
association
stage
association set
target
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201911036959.XA
Other languages
Chinese (zh)
Other versions
CN110686679A (en
Inventor
刘勇
曹璐
郭鹏宇
李献斌
王建
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
National Defense Technology Innovation Institute PLA Academy of Military Science
Original Assignee
National Defense Technology Innovation Institute PLA Academy of Military Science
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by National Defense Technology Innovation Institute PLA Academy of Military Science filed Critical National Defense Technology Innovation Institute PLA Academy of Military Science
Priority to CN201911036959.XA priority Critical patent/CN110686679B/en
Publication of CN110686679A publication Critical patent/CN110686679A/en
Application granted granted Critical
Publication of CN110686679B publication Critical patent/CN110686679B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/20—Instruments for performing navigational calculations
    • G01C21/203—Specially adapted for sailing ships
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Mathematical Physics (AREA)
  • Mathematical Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Computational Mathematics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Probability & Statistics with Applications (AREA)
  • Evolutionary Biology (AREA)
  • Operations Research (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Algebra (AREA)
  • Automation & Control Theory (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention discloses a high orbit optical satellite offshore target interruption track correlation method, which comprises the following steps: forward extrapolating the first track backward and the second track forward; performing primary association on a first track and a second track of which the state estimation at the same moment meets the constraint condition; further associating the flight paths in the preliminary association set to obtain a further association set; solving the further association set by adopting a two-dimensional optimal distribution principle to obtain a first-stage association set with the minimum global cost; solving possible association segment pairs by adopting distance two-dimensional optimal distribution on a complement set of a first-stage association set in the preliminary association set; detecting and screening the intermediate association set by adopting amplitude relation to obtain a second-stage association set; and fitting the associated track pairs in the first-stage associated set and the second-stage associated set to obtain a continuous track. The method for associating the interrupted flight path of the marine target of the high-orbit optical satellite effectively improves the performance of flight path association.

Description

High-orbit optical satellite offshore target interruption track correlation method
Technical Field
The invention relates to the technical field of space-based marine target monitoring, in particular to a high-orbit optical satellite marine target track interruption association method.
Background
The working modes of the high orbit (in particular to geostationary orbit) optical satellite can be different according to the observation task of the satellite, and take a high-resolution four-number (GF-4) satellite in China as an example, the working modes of the high orbit optical satellite are mainly divided into staring imaging, area imaging and maneuvering inspection. The staring imaging mode is used for continuously imaging a monitored area, the area imaging mode is used for splicing imaging of the monitored area, and the maneuvering inspection mode is used for quickly finishing imaging of a plurality of areas by using attitude maneuvering according to task requirements. Due to the restriction of satellite resources, the working time of the satellite every day is limited, and the satellite cannot be started for a long time. Therefore, the high-orbit optical satellite needs to acquire more spatial information by using less time resources, cover more target areas as much as possible, and realize wide-area search and tracking. In ocean surveillance, an optical satellite in high orbit can combine a staring imaging mode and a maneuvering inspection mode to continuously image a certain sea area at intervals. Because the track of the marine target can be formed in each staring imaging, the tracks at different imaging moments need to be correlated, namely, the track correlation is interrupted, the re-identification of the marine target is realized, and the marine situation perception capability of the area is improved. Meanwhile, whether the target deviates from an expected route can be judged by interrupting the track association, the target track with strong maneuverability is detected, and the abnormal track detection is realized. Through the correlated track connection, a more complete, continuous and accurate marine target track can be obtained. Therefore, it is necessary to provide a method for correlating the interrupted flight path of the marine target of the high-orbit optical satellite. The problem of association of broken track was proposed earlier in the eighties of the last century and has been a hot point of research by scholars at home and abroad. The traditional method for associating the interrupted flight path mainly solves the problem of flight path interruption caused by target shielding, target maneuvering and the like, and mainly performs flight path segment association based on statistical distance, so that the association effect is poor under the complex conditions of long time interval, dense target distribution and the like.
Disclosure of Invention
In order to solve the technical problems in the prior art, the invention provides a high-orbit optical satellite offshore target track interruption association method. The specific technical scheme is as follows:
a method for high-orbit optical satellite marine target outage trajectory correlation, the method comprising: backward extrapolating the first track and forward extrapolating the second track to obtain the state estimation of the first track and the second track at the same moment; performing primary association on a first flight path and a second flight path of which the state estimation at the same moment meets the constraint condition to obtain a primary association set; adopt χ for tracks in preliminary association set2Further association is carried out by the distributed hypothesis testing method, and a further association set is obtained; solving the further association set by adopting a two-dimensional optimal distribution principle to obtain an association set with the minimum global cost as a first-stage association set; solving possible association segment pairs by adopting distance two-dimensional optimal distribution on a complement of a first-stage association set in the preliminary association set to obtain a middle association set; screening the intermediate association set by adopting amplitude relation detection to obtain a second stage association set; and fitting the associated tracks in the first-stage associated set and the second-stage associated set by adopting an N-order polynomial to obtain a continuous track.
Optionally, the set of first tracks is:
Figure GDA0003032174170000021
i is the total number of first tracks,
Figure GDA0003032174170000022
respectively the starting time and the ending time of the first track; the set of second tracks is:
Figure GDA0003032174170000023
j is the total number of second tracks,
Figure GDA0003032174170000024
respectively the starting time and the ending time of the second track; the state estimation vector of the flight path at the moment k is as follows:
Figure GDA0003032174170000025
Figure GDA0003032174170000026
Figure GDA0003032174170000027
in the formula (I), the compound is shown in the specification,
Figure GDA0003032174170000028
respectively representing latitude, longitude, heading, and speed estimate of the target at time k,
all possible track combinations at time k are:
Figure GDA0003032174170000029
optionally, the preliminary association set is:
Figure GDA00030321741700000210
in the formula, sth,θth,pthRespectively, speed, course and position associated with the threshold, the position distance being approximately replaced by Euclidean distance, phivRepresenting a preliminary association set, Ti oA first of the flight paths is represented,
Figure GDA00030321741700000211
a second track is represented that is a second track,
Figure GDA00030321741700000212
indicating a first track at
Figure GDA00030321741700000213
The speed estimate at the time of day is,
Figure GDA00030321741700000214
show the second track at
Figure GDA00030321741700000215
The speed estimate at the time of day is,
Figure GDA00030321741700000216
indicating a first track at
Figure GDA00030321741700000217
The course estimate of the time of day is,
Figure GDA00030321741700000218
indicating a second track at
Figure GDA00030321741700000219
The course estimate of the time of day is,
Figure GDA00030321741700000220
is shown in
Figure GDA00030321741700000221
The difference in position between the second track and the first track at the time.
Alternatively, p is obtained by the following formulath：
Figure GDA0003032174170000031
Wherein p isth1,pth2Position changes, s, due to speed and direction changes, respectivelymeanIs the average velocity of the marine target.
Optionally, χ is adopted for the tracks in the preliminary association set2The distributed hypothesis testing method is further associated to obtain a further association set, and the method comprises the following steps: definition H0：
Figure GDA0003032174170000032
And
Figure GDA0003032174170000033
state estimation of a first track and a second track of the same target at the same time; at H0In the assumption that in the middle of the above,
Figure GDA0003032174170000034
first track of time Ti oAnd the second track
Figure GDA0003032174170000035
The estimation error is:
Figure GDA0003032174170000036
the covariance of the corresponding errors is:
Figure GDA0003032174170000037
then at a confidence of 1-Q, the set of further associations is:
Figure GDA0003032174170000038
wherein (D)ij)′(Pij)-1(Dij) Obey nxChi of degree of freedom2Distribution, nxIs the dimension of the state vector.
Optionally, solving the further association set by using a two-dimensional optimal allocation principle to obtain an association set with the minimum global cost, as a first-stage association set, includes:
performing global optimal association on the first track and the second track in the further association set, wherein an associated cost function is as follows:
Figure GDA0003032174170000039
Figure GDA00030321741700000310
the value of the two-dimensional distribution variable a (i, j) requires minimizing the weighted sum of the distribution costs, i.e.
Figure GDA00030321741700000311
Meanwhile, a (i, j) needs to satisfy the following constraint:
Figure GDA00030321741700000312
wherein, a (i, j) ═ {0,1}, a (i, j) ═ 1 denotes that two track segments are associated, otherwise, they are not related;
the two-dimensional distribution is finally solved through a Munkres algorithm to obtain a first-stage association set:
Figure GDA0003032174170000041
optionally, for a complement set of the first-stage association set in the preliminary association set, a distance two-dimensional optimal allocation is adopted to solve a possible association track pair through the following formula, so as to obtain an intermediate association set:
Figure GDA0003032174170000042
and is
Figure GDA0003032174170000043
Wherein, I 'is the number of the remaining first track segments, and J' is the number of the remaining second track segments.
Optionally, screening the intermediate association set by using amplitude relationship detection to obtain a second-stage association set, including:
analyzing the amplitude corresponding relation of the associated track pairs in the first-stage associated set through a linear regression model to obtain a standard deviation; and eliminating the associated track pairs with the amplitude prediction error larger than or equal to 3 times of the standard deviation in the intermediate associated set, and taking the residual set as a second-stage associated set.
Optionally, for the associated track pair already associated in the first stage, analyzing the amplitude correspondence between the first track and the second track through a linear regression model, including:
is provided withThe marine target amplitude set of the associated conditions is:
Figure GDA0003032174170000044
wherein the content of the first and second substances,
Figure GDA0003032174170000045
respectively obtaining the amplitudes of a first track and a second track in the ith associated track pair, wherein the amplitudes are the average value of the target amplitudes at each moment under current observation, and K is the number of the associated track pairs in the first-stage associated set;
the relationship between the target amplitudes under two observations is assumed to satisfy the linear regression:
Figure GDA0003032174170000046
and obtaining the amplitude of the associated track pair in the first stage association set by least square estimation:
Figure GDA0003032174170000047
with the coefficient determined by R2, the R2 expression is:
Figure GDA0003032174170000051
wherein SStotIs a sum of squares, SSresIs the sum of the squares of the residuals.
Optionally, for all the associated track pairs in the first-stage associated set and the second-stage associated set, performing the interrupted track connection by using the following formula:
Figure GDA0003032174170000052
and K' is the total number of the associated track pairs in the first-stage association set and the second-stage association set.
The technical scheme of the invention has the following main advantages:
according to the high-orbit optical satellite offshore target interruption track association method, firstly, the track segments meeting the constraint conditions are roughly associated, so that the candidate combinations of association are effectively reduced, and the association efficiency is improved. In the first stage of association, a two-dimensional optimal association combination is solved through motion information hypothesis test and cost function construction, and therefore the upper weak maneuvering target is associated. In the second stage of association, the two-dimensional optimal association combination is solved for the remaining track pairs by using the distance information, and the amplitude information is adopted for carrying out anomaly detection, so that the reliability and the confidence coefficient of track association are increased. And finally, fitting the associated track pair by adopting a polynomial to complete the connection and fusion of the tracks in the whole observation time. Under the complex conditions of long-time satellite interval, dense target distribution and the like, the track association accuracy is high, the robustness is strong, the continuous monitoring capability of the high-orbit satellite on the marine target can be effectively improved, and the track association capability is effectively improved.
Drawings
The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of a method for associating a broken track with an offshore target of an optical satellite in high orbit according to an embodiment of the present invention;
FIG. 2 is a diagram of a track association framework provided by one embodiment of the present invention;
FIG. 3 is a flow chart of the interrupt track association provided by one embodiment of the present invention;
FIG. 4-1 is a scene diagram of a track-associated scene 1 according to an embodiment of the present invention;
FIG. 4-2 is a scene diagram of a track-associated scene 2 according to an embodiment of the present invention;
FIG. 5-1 is a schematic diagram illustrating a first-stage track-break association of scenario 1 according to an embodiment of the present invention;
FIG. 5-2 is a schematic diagram illustrating a first-stage track-break association of scenario 2 according to an embodiment of the present invention;
FIG. 6-1 is a schematic diagram illustrating a second phase interrupted track correlation of scenario 1 according to an embodiment of the present invention;
FIG. 6-2 is a schematic diagram of a second phase interrupted track correlation of scenario 2 according to an embodiment of the present invention;
FIG. 7-1 is a schematic diagram of an interrupted track link of scenario 1 according to an embodiment of the present invention;
fig. 7-2 is a schematic view of broken track connection of scenario 2 according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the specific embodiments of the present invention and the accompanying drawings. It is to be understood that the described embodiments are merely a few embodiments of the invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
The technical scheme provided by the embodiment of the invention is described in detail below with reference to the accompanying drawings.
The embodiment of the invention provides a method for associating a broken track of an offshore target of a high-orbit optical satellite, which comprises the following steps of as shown in attached figures 1 and 3:
and S1, backward extrapolating the first track and forward extrapolating the second track to obtain the state estimation of the first track and the second track at the same time.
And S2, performing preliminary association on the first track and the second track of which the state estimation at the same time meets the constraint condition to obtain a preliminary association set. Corresponding to the track coarse correlation step in fig. 3.
S3, adopting χ for tracks in the preliminary association set2And further associating the distributed hypothesis testing methods to obtain a further association set. Corresponding to the hypothesis testing step in the first stage correlation of fig. 3.
And S4, solving the further association set by adopting a two-dimensional optimal distribution principle to obtain an association set with the minimum global cost as a first-stage association set. Corresponding to the two-dimensional optimal allocation step in the first stage association of fig. 3.
And S5, solving possible association segment pairs by adopting distance two-dimensional optimal distribution on the complement of the first-stage association set in the preliminary association set to obtain a middle association set. Corresponding to the two-dimensional optimal allocation step in the second stage association of FIG. 3
And S6, screening the intermediate association set by adopting amplitude relation detection to obtain a second stage association set. Corresponding to the step of detecting the amplitude relationship in the second stage of association in fig. 3.
And S7, fitting the associated tracks in the first-stage associated set and the second-stage associated set by adopting an N-order polynomial to obtain continuous tracks. Corresponding to the track segment fitting and fused track output in figure 3.
According to the method for associating the interrupted flight path of the high orbit optical satellite marine target, provided by the embodiment of the invention, the flight path segments meeting the constraint condition are roughly associated, so that the candidate combination of association is effectively reduced, and the association efficiency is improved. In the first stage of association, a two-dimensional optimal association combination is solved through motion information hypothesis test and cost function construction, and therefore the upper weak maneuvering target is associated. In the second stage of association, the two-dimensional optimal association combination is solved for the remaining track pairs by using the distance information, and the amplitude information is adopted for carrying out anomaly detection, so that the reliability and the confidence coefficient of track association are increased. And finally, fitting the associated track pair by adopting a polynomial to complete the connection and fusion of the tracks in the whole observation time. Under the complex conditions of long-time satellite interval, dense target distribution and the like, the track association accuracy is high, the robustness is strong, the continuous monitoring capability of the high-orbit satellite on the marine target can be effectively improved, and the track association capability is effectively improved.
The overall frame of the track association can be seen in fig. 2, a high-orbit optical satellite (taking GF-4 as an example) obtains two satellite picture sequences through two successive observations, ship detection and tracking are performed in the two satellite picture sequences, a first track and a second track can be respectively obtained, the first track and the second track are associated and fused into a continuous track, and a report is generated and output outwards.
Each flow in the method for associating the interrupted flight path of the marine target of the high orbit optical satellite provided by the embodiment of the invention is explained in detail as follows:
the invention takes GF-4 satellite as an example for explanation, and GF-4 satellite images adopt static correction and target detection and tracking methods. Assume that at an interval time Δ T, the high-orbit optical satellite images the monitored sea area twice, each monitoring lasting for a period of time, generating a track segment of the marine target. Interrupting the track association requires determining whether the tracks of two previous and subsequent observations are the tracks of the same marine target, wherein the track segments of the two observations are defined as a first track (i.e., the first observed track) and a second track (i.e., the second observed track).
The set of first tracks is:
Figure GDA0003032174170000071
i is the total number of first tracks,
Figure GDA0003032174170000072
respectively the starting time and the ending time of the first track;
the set of second tracks is:
Figure GDA0003032174170000073
j is the total number of second tracks,
Figure GDA0003032174170000074
respectively the start and end times of the second track.
The state estimation vector of the flight path at the moment k is as follows:
Figure GDA0003032174170000075
Figure GDA0003032174170000076
Figure GDA0003032174170000081
in the formula (I), the compound is shown in the specification,
Figure GDA0003032174170000082
respectively representing latitude, longitude, heading, and speed estimate of the target at time k.
Given the parameters in the state estimation vector of the track, it can be understood that, in the following appearing parameters, the parameter in the state estimation vector representing the first track is added with the superscript i, and the parameter in the state estimation vector representing the second track is added with the superscript j.
All possible track combinations at time k are:
Figure GDA0003032174170000083
by backward extrapolation of the first track and forward extrapolation of the new track, state estimates at k times of the first track and the second track can be obtained, which are respectively expressed as
Figure GDA0003032174170000084
And
Figure GDA0003032174170000085
then is at
Figure GDA0003032174170000086
The state estimates at the time are respectively
Figure GDA0003032174170000087
And
Figure GDA0003032174170000088
suppose a first track Ti oAnd the second track
Figure GDA0003032174170000089
Two tracks of the same marine target are required to meet certain constraint conditions in speed, direction and distance. By setting a coarse approximationAnd (3) bundling conditions to obtain a preliminary association set, wherein the preliminary association set is as follows:
Figure GDA00030321741700000810
in the formula phivRepresenting a preliminary association set, Ti oA first of the flight paths is represented,
Figure GDA00030321741700000811
a second track is represented that is a second track,
Figure GDA00030321741700000812
indicating a first track at
Figure GDA00030321741700000813
The speed estimate at the time of day is,
Figure GDA00030321741700000814
show the second track at
Figure GDA00030321741700000815
The speed estimate at the time of day is,
Figure GDA00030321741700000816
indicating a first track at
Figure GDA00030321741700000817
The course estimate of the time of day is,
Figure GDA00030321741700000818
indicating a second track at
Figure GDA00030321741700000819
The course estimate of the time of day is,
Figure GDA00030321741700000820
is shown in
Figure GDA00030321741700000821
The difference in position between the second track and the first track at the time.
sth,θth,pthRespectively, speed, heading and position are associated with the threshold, and the position distance is approximately replaced by Euclidean distance. Position deviation (difference between actual position and predicted position) is caused by factors such as speed, heading, and interruption time interval. In the embodiment of the present invention, the setting mode of the position association threshold is as follows:
Figure GDA00030321741700000822
wherein p isth1,pth2Position changes, s, due to speed and direction changes, respectivelymeanIs the average velocity of the marine target. In practical application, corresponding threshold values need to be set according to the maneuvering condition of the marine target in the monitoring area. For example, s is set in the embodiment of the present inventionth、smean20kn and 10kn, thetathIs 60 degree
After the track rough association is completed, because the track state estimation errors of the same target at the same moment are independent in statistics and obey Gaussian distribution, Chi can be adopted2The distributed hypothesis testing method further correlates the track segments. Let H0,H1Two assumptions, H, respectively indicating whether track segments are related or not0,H1Is defined as:
H0：
Figure GDA0003032174170000091
and
Figure GDA0003032174170000092
the state estimation of new and old tracks of the same target at the same moment;
H1：
Figure GDA0003032174170000093
and
Figure GDA0003032174170000094
new and old flight paths not belonging to the same target
At H0In the assumption that in the middle of the above,
Figure GDA0003032174170000095
old track of time Ti oAnd new flight path
Figure GDA0003032174170000096
The estimation error is:
Figure GDA0003032174170000097
the covariance of the corresponding errors is:
Figure GDA0003032174170000098
then at a confidence of 1-Q, the set of further associations is:
Figure GDA0003032174170000099
wherein (D)ij)′(Pij)-1(Dij) Obey nxChi of degree of freedom2Distribution, nxIs the dimension of the state vector.
After the hypothesis test is completed, performing global optimal correlation on the first track and the second track in the further correlation set, wherein the correlated cost function is the estimation error DijA likelihood function of (a);
Figure GDA00030321741700000910
Figure GDA00030321741700000911
when performing two-dimensional optimal allocation, the value of the two-dimensional allocation variable a (i, j) needs to minimize the weighted sum of the allocation costs, i.e. the weighted sum
Figure GDA00030321741700000912
Meanwhile, a (i, j) needs to satisfy the following constraint:
Figure GDA00030321741700000913
where a (i, j) {0,1}, a (i, j) } 1 indicates that two track segments are associated, and otherwise, are not related. It can be seen from the above two formulas that each second track can be associated with only one first track at most, and each first track can be associated with only one second track at most. The two-dimensional distribution is finally solved through a Munkres algorithm to obtain a first-stage association set:
Figure GDA0003032174170000101
at this point, the first stage of association is completed. In the first stage of association, hypothesis test is carried out on the motion state of the target, and the method belongs to a conservative association strategy, so that the reliability of track segment association can be ensured. However, under a long time interval, part of marine targets can show some maneuverability, such as turning, changing direction, accelerating and the like, if only the motion state is used for correlating the rest tracks, some tracks can be wrongly correlated, and therefore the motion information and the amplitude information of the targets are utilized to perform second-stage correlation.
Specifically, for the complement of the first-stage association set in the preliminary association set, a distance two-dimensional optimal allocation is adopted to solve possible association track pairs by the following formula, and a middle association set is obtained:
Figure GDA0003032174170000102
and is
Figure GDA0003032174170000103
Wherein, I 'is the number of the remaining first track segments, and J' is the number of the remaining second track segments.
Further, in order to improve the confidence of the track association, the intermediate association set is screened by adopting amplitude relationship detection, and a second-stage association set is obtained, including:
analyzing the amplitude corresponding relation of the associated track pairs in the first-stage associated set through a linear regression model to obtain a standard deviation; and eliminating the associated track pairs with the amplitude prediction error larger than or equal to 3 times of the standard deviation in the intermediate associated set, and taking the residual set as a second-stage associated set.
For the amplitude relationship solution, the following is elaborated:
under the influence of observation time, the conditions such as illumination and the like in two times of observation are different, so that the target amplitude of the same target in the new and old tracks is different, but has certain correlation. Ideally, the amplitude of the image, i.e. the relative radiation correction, can be adjusted by means of image histogram registration, so that the invariance of the amplitude of the same marine target in two observations is realized. However, the histogram matching is an integral matching, and the sea surface background occupies a larger proportion of the image, so that the sea object occupies fewer pixels, which causes non-uniformity during histogram matching and inaccurate sea object amplitude correction. Therefore, the invention directly utilizes the previously associated marine target pairs and adopts a linear regression model to analyze the corresponding relation of the amplitude of the marine target observed twice. The marine target amplitude set meeting the correlation condition is set as follows:
Figure GDA0003032174170000104
wherein the content of the first and second substances,
Figure GDA0003032174170000111
respectively obtaining the amplitudes of a first track and a second track in the ith associated track pair, wherein the amplitudes are the average value of the target amplitudes at each moment under current observation, and K is the number of the associated track pairs in the first-stage associated set;
the relationship between the target amplitudes under two observations is assumed to satisfy the linear regression:
Figure GDA0003032174170000112
and obtaining the amplitude of the associated track pair in the first stage association set by least square estimation:
Figure GDA0003032174170000113
by the use of R2Determination coefficient, R2The expression is as follows:
Figure GDA0003032174170000114
wherein SStotIs a sum of squares, SSresIs the sum of the squares of the residuals.
After the first-stage association and the second-stage association are completed, in order to improve the integrity and continuity of tracks in the monitoring area, track connection needs to be performed on the basis of track segment association. Wherein, the fitting data adopts the position state estimation value in the flight path segment, namely:
Figure GDA0003032174170000115
and K' is the total number of the associated track pairs in the first-stage association set and the second-stage association set. Between observations of a GF-4 satellite, the motion state of the marine target may change, causing the marine target position to deviate from equiangular course travel. In order to smoothly connect the first track and the second track, the estimated values of the position states of the longitude and latitude direction axes are respectively fitted by adopting an N-order polynomial by taking time as an independent variable, so that the track of the target in the whole time period is obtained. When the fitting order is larger, the fitting curve has larger volatility and is not consistent with the target motion state; when the fitting order is small, such as 1 st order straight line fitting, it is difficult to accurately describe a complex motion state. According to the general motion situation of the sea object in practice, 3-order fitting is selected in the embodiment of the invention.
The method for associating the interrupted flight path of the marine target of the high-orbit optical satellite is further described by combining specific embodiments in the following steps:
in order to verify the effectiveness of the high-orbit optical satellite interrupted track correlation method, two GF-4 satellite imaging scenes in the east sea area are selected, as shown in FIG. 4, and "Coastline", "GF-4T 1", "GF-4T 2" and "ROI" respectively represent Coastline data, a first imaging area, a second imaging area and a region of interest (ROI) selected by research. The interruption time of scene 1 is about 1 hour, and the time interval of scene 2 is about 2 times of scene 1. And the track of the marine target can be obtained by detecting and tracking the marine target. Compared with scenario 1, scenario 2 has more false alarm targets and a greater number of real targets, and the correlation difficulty is greater when the satellite interval is longer.
Fig. 5-1 and 5-2 are the correlation results of the motion information in the first stage under the scene 1 and the scene 2, respectively, and the weak-mobility marine target can be correlated by performing hypothesis testing and two-dimensional optimal distribution by using the motion information. Since the target on the correlation needs to follow the approximately straight-line navigation and is the correlation under the strong condition, the correlation result has high reliability.
Fig. 6-1 and 6-2 are flight path association diagrams using motion information and amplitude information in the second stage of the scene 1 and the scene 2, and it can be seen that the target of the upper motion deviating from the predicted flight path can be effectively associated by using the interrupted flight path association of the amplitude information. In two scenes, the performance of the traditional interrupted track correlation method is equivalent to that of the method of the invention on the track correlation accuracy, and is close to 100%. However, on the track association integrity rate, the method of the invention is improved to 100% from 84% of the traditional method, so that more tracks can be correctly associated, and the performance of interrupting track association is effectively improved.
Fig. 7-1 and 7-2 are diagrams of the effect of breaking track connection of scene 1 and scene 2, respectively, and it can be seen that the track connection can grasp the track dynamics in the whole monitoring time and region, and the 3 rd order polynomial fitting adopted by the invention is more in line with the motion track of the marine target. The movement of the part of the marine target has mobility, the final position deviates from the predicted position due to the change of the speed or the direction, and the track of the mobile target can be accurately mastered through track association and connection.
In the embodiment, each imaging of each selected scene is 4 images, and the imaging interval is 1-2.5 hours. As can be seen, the motion perception of the marine target for a long time can be realized only by multi-frame data observed twice in the same area in a track segment association mode, and satellite resources are greatly saved.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. In addition, "front", "rear", "left", "right", "upper" and "lower" in this document are referred to the placement states shown in the drawings.
Finally, it should be noted that: the above examples are only for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (8)

1. A method for high-orbit optical satellite offshore target outage trajectory correlation, the method comprising:
backward extrapolating the first track and forward extrapolating the second track to obtain the state estimation of the first track and the second track at the same moment;
performing primary association on a first flight path and a second flight path of which the state estimation at the same moment meets the constraint condition to obtain a primary association set;
adopt χ for tracks in preliminary association set2Further association is carried out by the distributed hypothesis testing method, and a further association set is obtained;
solving the further association set by adopting a two-dimensional optimal distribution principle to obtain an association set with the minimum global cost as a first-stage association set;
solving possible associated track pairs by adopting distance two-dimensional optimal distribution on a complement of a first-stage associated set in the preliminary associated set to obtain a middle associated set;
screening the intermediate association set by adopting amplitude relation detection to obtain a second stage association set;
fitting the associated tracks in the first-stage associated set and the second-stage associated set by adopting an N-order polynomial to obtain continuous tracks;
the screening of the intermediate association set by adopting amplitude relation detection to obtain the second stage association set comprises the following steps:
analyzing the amplitude corresponding relation of the associated track pairs in the first-stage associated set through a linear regression model to obtain a standard deviation;
removing the associated track pairs with the amplitude prediction error larger than or equal to 3 times of the standard deviation in the intermediate associated set, and taking the residual set as a second-stage associated set;
for the associated track pair associated in the first stage, analyzing the amplitude corresponding relation of the first track and the second track through a linear regression model, wherein the amplitude corresponding relation comprises the following steps:
the marine target amplitude set meeting the correlation condition is set as follows:
Figure FDA0003032174160000011
wherein the content of the first and second substances,
Figure FDA0003032174160000012
respectively obtaining the amplitudes of a first track and a second track in the ith associated track pair, wherein the amplitudes are the average value of the target amplitudes at each moment under current observation, and K is the number of the associated track pairs in the first-stage associated set;
the relationship between the target amplitudes under two observations is assumed to satisfy the linear regression:
Figure FDA0003032174160000013
and obtaining the amplitude of the associated track pair in the first stage association set by least square estimation:
Figure FDA0003032174160000014
by the use of R2Determination coefficient, R2The expression is as follows:
Figure FDA0003032174160000021
wherein SStotIs a sum of squares, SSresIs the sum of the squares of the residuals.
2. The method of claim 1, wherein the first set of tracks is:
Figure FDA0003032174160000022
i is the sum of the first trackThe number of the first and second groups is,
Figure FDA0003032174160000023
respectively the starting time and the ending time of the first track;
the set of second tracks is:
Figure FDA0003032174160000024
j is the total number of second tracks,
Figure FDA0003032174160000025
respectively the starting time and the ending time of the second track;
the state estimation vector of the flight path at the moment k is as follows:
Figure FDA0003032174160000026
Figure FDA0003032174160000027
Figure FDA0003032174160000028
in the formula (I), the compound is shown in the specification,
Figure FDA0003032174160000029
respectively representing the latitude, longitude, course and speed estimation of the target at the moment k, wherein all possible tracks at the moment k are combined as follows:
Figure FDA00030321741600000210
and increasing the superscript i to represent the parameters in the state estimation vector of the first track, and increasing the superscript j to represent the parameters in the state estimation vector of the second track.
3. The method of claim 2, wherein the preliminary association set is:
Figure FDA0003032174160000031
in the formula, sth,θth,pthRespectively, speed, course and position associated with the threshold, the position distance being approximately replaced by Euclidean distance, phivRepresenting a preliminary association set, Ti oA first of the flight paths is represented,
Figure FDA0003032174160000032
a second track is represented that is a second track,
Figure FDA0003032174160000033
indicating a first track at
Figure FDA0003032174160000034
The speed estimate at the time of day is,
Figure FDA0003032174160000035
show the second track at
Figure FDA0003032174160000036
The speed estimate at the time of day is,
Figure FDA0003032174160000037
indicating a first track at
Figure FDA0003032174160000038
The course estimate of the time of day is,
Figure FDA0003032174160000039
indicating a second track at
Figure FDA00030321741600000310
The course estimate of the time of day is,
Figure FDA00030321741600000311
is shown in
Figure FDA00030321741600000312
The difference in position between the second track and the first track at the time.
4. The method as claimed in claim 3, wherein p is obtained by the following formulath：
Figure FDA00030321741600000313
Wherein p isth1,pth2Position changes, s, due to speed and direction changes, respectivelymeanΔ T represents the time interval between two imaging, which is the average velocity of the marine target.
5. The method as claimed in claim 3, wherein χ is adopted for the tracks in the preliminary association set2The distributed hypothesis testing method is further associated to obtain a further association set, and the method comprises the following steps:
definition H0：
Figure FDA00030321741600000314
And
Figure FDA00030321741600000315
state estimation of a first track and a second track of the same target at the same time;
at H0In the assumption that in the middle of the above,
Figure FDA00030321741600000316
first track of time Ti oAnd the second track
Figure FDA00030321741600000317
The estimation error is:
Figure FDA00030321741600000318
the covariance of the corresponding errors is:
Figure FDA00030321741600000319
then at a confidence of 1-Q, the set of further associations is:
Figure FDA0003032174160000041
wherein (D)ij)′(Pij)-1(Dij) Obey nxChi of degree of freedom2Distribution, nxIs the dimension of the state vector.
6. The method for associating interrupted tracks of offshore targets with the high-orbit optical satellite according to claim 5, wherein the further association set is solved by adopting a two-dimensional optimal distribution principle to obtain an association set with the minimum global cost, and the method for associating interrupted tracks of offshore targets with the high-orbit optical satellite as the first-stage association set comprises the following steps:
performing global optimal association on the first track and the second track in the further association set, wherein an associated cost function is as follows:
Figure FDA0003032174160000042
Figure FDA0003032174160000043
the value of the two-dimensional distribution variable a (i, j) requires minimizing the weighted sum of the distribution costs, i.e.
Figure FDA0003032174160000044
Meanwhile, a (i, j) needs to satisfy the following constraint:
Figure FDA0003032174160000045
wherein, a (i, j) ═ {0,1}, a (i, j) ═ 1 denotes that two track segments are associated, otherwise, they are not related;
the two-dimensional distribution is finally solved through a Munkres algorithm to obtain a first-stage association set:
Figure FDA0003032174160000046
7. the method for high-orbit optical satellite offshore target interrupted track correlation according to claim 6, wherein the possible associated track pairs are solved for the complement set of the first-stage associated set in the preliminary associated set by adopting distance two-dimensional optimal distribution through the following formula to obtain a middle associated set:
Figure FDA0003032174160000047
and is
Figure FDA0003032174160000048
Wherein, I 'is the number of the remaining first track segments, and J' is the number of the remaining second track segments.
8. The off-track correlation method for the marine target of the high-orbit optical satellite according to claim 7, wherein for all the associated track pairs in the first-stage association set and the second-stage association set, the off-track connection is performed by polynomial fitting, and the polynomial fitting data is estimated by using the position state in the track segment, that is:
Figure FDA0003032174160000051
and K' is the total number of the associated track pairs in the first-stage association set and the second-stage association set.
CN201911036959.XA 2019-10-29 2019-10-29 High-orbit optical satellite offshore target interruption track correlation method Active CN110686679B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911036959.XA CN110686679B (en) 2019-10-29 2019-10-29 High-orbit optical satellite offshore target interruption track correlation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911036959.XA CN110686679B (en) 2019-10-29 2019-10-29 High-orbit optical satellite offshore target interruption track correlation method

Publications (2)

Publication Number Publication Date
CN110686679A CN110686679A (en) 2020-01-14
CN110686679B true CN110686679B (en) 2021-07-09

Family

ID=69114557

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911036959.XA Active CN110686679B (en) 2019-10-29 2019-10-29 High-orbit optical satellite offshore target interruption track correlation method

Country Status (1)

Country Link
CN (1) CN110686679B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112686095B (en) * 2020-12-04 2022-08-19 中国人民解放军海军航空大学 Ship target track correlation method for stationary orbit staring satellite remote sensing image
CN113516037B (en) * 2021-05-11 2024-01-23 中国石油大学(华东) Marine vessel track segment association method, system, storage medium and equipment
CN115525640B (en) * 2022-11-25 2023-04-07 三亚海兰寰宇海洋信息科技有限公司 Target object trajectory processing method, device and equipment

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104318094A (en) * 2014-10-15 2015-01-28 中国电子科技集团公司第二十八研究所 Association method for radar networking flight path for formation flight
CN105843919A (en) * 2016-03-24 2016-08-10 云南大学 Moving object track clustering method based on multi-feature fusion and clustering ensemble
CN108168564A (en) * 2017-12-04 2018-06-15 上海无线电设备研究所 A kind of Data Association based on LHD grey relational grades
CN109085569A (en) * 2018-08-13 2018-12-25 南京邮电大学 A kind of more radar track correlating methods based on region division

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104964690A (en) * 2015-05-22 2015-10-07 西北工业大学 Expectation-maximization-based aerial maneuvering target track segment association method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104318094A (en) * 2014-10-15 2015-01-28 中国电子科技集团公司第二十八研究所 Association method for radar networking flight path for formation flight
CN105843919A (en) * 2016-03-24 2016-08-10 云南大学 Moving object track clustering method based on multi-feature fusion and clustering ensemble
CN108168564A (en) * 2017-12-04 2018-06-15 上海无线电设备研究所 A kind of Data Association based on LHD grey relational grades
CN109085569A (en) * 2018-08-13 2018-12-25 南京邮电大学 A kind of more radar track correlating methods based on region division

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
俞建国等.弹道目标航迹片段关联及优化.《航空学报》.2011,第32卷(第10期),第1897-1903页. *

Also Published As

Publication number Publication date
CN110686679A (en) 2020-01-14

Similar Documents

Publication Publication Date Title
CN110686679B (en) High-orbit optical satellite offshore target interruption track correlation method
CN102881024B (en) Tracking-learning-detection (TLD)-based video object tracking method
TWI393074B (en) Apparatus and method for moving object detection
CN104008371B (en) Regional suspicious target tracking and recognizing method based on multiple cameras
CN101511022B (en) Method for implementing machine-carried video compression and target tracking unitedly
Bloisi et al. Argos—A video surveillance system for boat traffic monitoring in Venice
CN102456226B (en) Region-of-interest tracking system, method and computer program product
WO2012009947A1 (en) Device and method of signature-driven multi-target tracking
CN110390292A (en) Based on the remote sensing video frequency vehicle object detecting and tracking method for dynamically associating model
CN112686095B (en) Ship target track correlation method for stationary orbit staring satellite remote sensing image
CN102609945B (en) Automatic registration method of visible light and thermal infrared image sequences
CN112488061B (en) Multi-aircraft detection and tracking method combined with ADS-B information
CN112346096B (en) High-low orbit remote sensing satellite ship target track and point track correlation method and system
CN110189363B (en) Airport scene moving target low-visual-angle video speed measuring method
CN112649802B (en) Tracking method before weak and small multi-target detection of high-resolution sensor
CN105825520A (en) Monocular SLAM (Simultaneous Localization and Mapping) method capable of creating large-scale map
CN111160212A (en) Improved tracking learning detection system and method based on YOLOv3-Tiny
CN108334885A (en) A kind of video satellite image space object detection method
CN111968046A (en) Radar photoelectric sensor target association fusion method based on topological structure
CN110458089B (en) Marine target association system and method based on high-low orbit optical satellite observation
CN114280611A (en) Road side sensing method integrating millimeter wave radar and camera
Wofk et al. Monocular visual-inertial depth estimation
KR100994722B1 (en) Method for tracking moving object on multiple cameras using probabilistic camera hand-off
CN105869108B (en) A kind of method for registering images in the mobile target detecting of moving platform
Bloisi et al. A distributed vision system for boat traffic monitoring in the venice grand canal.

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant